Most HIV-infected patients who stop taking antiretroviral drugs experience rapid viral rebound and suffer in the longer term from higher viral load and lower CD4+ T cell counts. This outcome normally occurs even when individuals have been suppressed for long periods, e.g., several decades. Thus, slow or non-existent viral rebound may occur only under restricted circumstances, including (i) early initiation of antiretroviral therapy, resulting in predominant infection of reservoir cells with shorter half-life, (ii) change in the viral reservoir under ART coverage, e.g., due to myeloablation in preparation for BMT, or (iii) change in the host's capacity for adaptive immune control over virus, e.g., due to therapeutic immunization. The relationships between time of ART initiation, duration of therapy, host immune responses, and kinetics of viral rebound are largely unknown. Given intriguing data indicating that certain host immune responses can contain nascent viral replication and even eliminate virus after low-dose SIV challenge, we propose a program of research on how host T cell responses might impact upon rebounding virus and/or the reservoir from which it springs. Prior studies have characterized rebound as a stochastic process that is most dependent on the number of reservoir cells remaining after ART cessation. We recognize the relevance of this factor but suggest that other reservoir characteristics are important. First, the activity of the viral reservoir, as assessed by cell-associated viral RNA, has recently been demonstrated to be a correlate of more rapid rebound. This finding is interesting because (i) it stimulates one to question if the latent and active reservoirs are separate pools, e.g., due to different viral integration sites, or a single pool of cells that stochastically re-activate transcription and because (ii) cells that are transcribing viral RNA might be vulnerable to host immune control. Second, the anatomic location of the viral reservoir might be important: it appears that RhCMV/SIV can contain viral replication to the portal of entry but have little impact following dissemination; thus, the same vaccines might contain spread of virus from a restricted reservoir. Finally, the phenotype of host cells harboring the reservoir would seem crucially important as their lifespan, anatomic location, vulnerability to antiretroviral drugs, interaction with innate or adaptive immunity, and indeed the kind of released virus will all differ. The objectives of our proposed program are: (i) to understand if host immunity impacts upon rebound via transformative effects on the reservoir itself, especially elimination of the active T cell reservoir; (ii) to understand if host immune responses can impact directly on viral replication or post-interruption setpoint in ways reminiscent of their demonstrated effects in primary SIV infection; and (iii) to test if HLA-E-restricted T cells have a qualitatively unique influence on viral rebound, as compared to effects of classical class Ia-restricted T cells or other host immune responses.